Dengue remains a significant global health burden, affecting 390 million people annually1, contributing to an estimated $8.9 billion in economic losses2. The dengue virus (DENV) exists as four distinct serotypes: DENV-1, DENV-2, DENV-3 and DENV-4. A major complication occurs during secondary infections, where cross-reactive antibodies from prior infection facilitate viral entry into immune cells, leading to antibody-dependent enhancement (ADE) and severe disease3. This phenomenon presents a major hurdle for dengue vaccine development, as effective dengue vaccines must protect against all four DENV serotypes, while also mitigating the risk of ADE. Adding to the complexity is the genetic variability within each serotype, with multiple genotypes circulating globally. Protection against one genotype may not confer lasting immunity to others4,5,6, complicating the long-term efficacy of dengue vaccines.

ADE is recognized as a critical factor in severe dengue, yet current dengue vaccine trials often focus primarily on neutralizing antibody (nAb) titers as a marker of protection. This approach was notably challenged by Dengvaxia®, the first licensed dengue vaccine, which induced adequate nAb titers but increased the risk of severe dengue in dengue-naïve individuals7, highlighting that nAb titers alone may not be predictive of vaccine efficacy8. Our group suggests that dengue vaccine development strategies should account for both neutralizing and ADE responses to ensure safety across populations with diverse immune backgrounds.

Live-attenuated tetravalent vaccines (LATVs) represent a promising solution by inducing humoral and cellular immunity across all four serotypes, closely mimicking natural infection. Among LATV candidates, KD-382 developed by KM Biologics shows potential as a single-dose dengue vaccine. KD-382 incorporates all four DENV serotypes, each attenuated by classic host-range mutations and temperature-sensitive mutations, which enables it to induce protective immune responses against both structural and nonstructural (NS) proteins from all four DENV serotypes9,10,11. However, achieving broad efficacy requires rigorous testing against the genetic diversity of circulating DENV strains, as genotypic variations can influence vaccine-induced immunity12.

This exploratory study utilized samples from the Phase I clinical trial of KD-382 to evaluate genotype-specific neutralization and potential in vitro ADE risk. We analyzed serum samples from 15 flavivirus-naïve healthy adults who received a single low dose of KD-382 (3 Log10 FFU per serotype), characterizing neutralizing and enhancing antibody responses over a 12-month period at months 1, 2, 3, 6 and 12. To assess the breadth of vaccine-induced immunity, we employed single-round infectious particles (SRIPs) representing 17 genetically diverse DENV genotypes from all four DENV serotypes. By focusing on individuals with minimal antigen exposure, this study offers valuable insights into the durability of vaccine-induced immune response and potential ADE risk in a lower-immunity context. We believe that this group is particularly informative for evaluating antibody waning and the possible emergence of infection-enhancing antibody profiles—critical factors for understanding whether KD-382 could inadvertently enhance infection rather than provide protection. These evaluations are especially relevant given the absence of standardized assays and the limited integration of ADE risk assessment in current dengue vaccine development strategies.

Our first step was to analyze the DENV-specific IgM and IgG levels in the 15 healthy, flavivirus-naïve adults who received a single low dose of KD-382. We used in-house ELISA systems to measure IgM13 and IgG14 titers over a 12-month period. We observed a robust DENV-specific IgM response after vaccination, which peaked shortly after vaccination and declined within three months (geometric mean P/N ratio: 5.0, 95% CI: 4.7–6.8). However, low but detectable levels of IgM persisted at 12 months (geometric mean P/N ratio: 2.5, 95% CI: 1.9–3.2) (Fig. 1a), suggesting ongoing, low-level immune surveillance. In contrast, DENV-specific IgG responses demonstrated long-term stability, peaking at two months (geometric mean IgG titers: 29,731.0 ± 1.4) with no significant fluctuations (Fig. 1b) throughout the 12-month observation period, supporting the potential for long-lasting immunity, particularly in regions with continuous DENV exposure.

Fig. 1: Longitudinal analysis of DENV-specific IgM and IgG responses over 12 months following single low-dose KD-382 administration.
figure 1

DENV-specific (a) IgM and (b) IgG levels were measured over a 12-month period following a single low-dose administration of KD-382. Statistical significance was evaluated using the non-parametric Kruskal–Wallis test, followed by Dunn’s multiple comparisons test comparing each time point (months 2 to 12) to month 1 as the reference. Dashed lines represent assay cut-off values for seropositivity: a P/N ratio ≥2.0 for IgM and a titer of 3000 for IgG.

Full size image

To assess the breadth and durability of the nAb response induced by a single low dose of KD-382, we analyzed serum samples for nAb titers against all four DENV serotypes over a 12-month observation period. Here, we employed a BHK-21 cell-based neutralization assay using DENV SRIPs to evaluate antibody responses targeting the structural surface proteins prM and E. The E protein is the key target of type-specific neutralizing antibodies, while the prM protein is associated with cross-reactive antibodies that may contribute to ADE of DENV infection15. The SRIP system used in this study incorporates a standardized yellow fever virus (YFV) replicon backbone16 across genotypes, allowing a focused assessment of nAbs directed specifically at prM and E. At 12 months post-vaccination, a single low dose of KD-382 elicited robust and durable nAb responses against DENV-1, DENV-2 and DENV-4, with geometric mean titers (GMTs; 95% confidence intervals) of 120 (86–167), 128 (83–198), and 88 (50–153), respectively (Table 1). Although there is no universally accepted correlate of protection for dengue, a nAb titer of ≥10 is commonly used as a benchmark for seropositivity in flavivirus vaccines, including those for Japanese encephalitis virus (JEV)17 and YFV. Based on this reference, KD-382 achieved seropositivity against all four DENV serotypes, including DENV-3, which demonstrated a lower GMT of 24 (17–34) (Table 1). Across all time points, GMTs against DENV-1 remained relatively stable, while DENV-2 and DENV-4 exhibited peak GMTs at one to two months post-vaccination followed by gradual declines (Table 1)—patterns consistent with those observed in the KD-382 Phase I clinical trial using the focus reduction neutralization test (FRNT)18. In contrast, GMTs against DENV-3 were consistently lower than those for the other serotypes throughout the 12-month observation period (Table 1), also aligning with the KD-382 Phase I clinical trial18. Despite the lower magnitude, nAb titers to DENV-3 were sustained over 12 months. Furthermore, evidence from the KD-382 non-human primate (NHP) study demonstrated that nAb titers to DENV-3 persisted up to 5 years10, comparable to the responses seen with the other DENV serotypes. These findings collectively support the durability of KD-382-induced nAb response against DENV-1, DENV-2, and DENV-4, and suggest that the nAb response to DENV-3, while lower, may still contribute to long-term protection, warranting continued follow-up in clinical settings.

Table 1 Serotype-specific GMTs and seroconversion to multiple serotypes in flavivirus-naïve healthy individuals following a single low-dose administration of KD-382
Full size table

Results from the KD-382 Phase I clinical trial18 showed that by month 2, 100% of participants achieved seropositivity (defined as FRNT50 ≥ 10) for all four DENV serotypes. At month 1, however, seropositivity was 100% for DENV-1 and DENV-2, while slightly lower rates were observed for DENV-3 (91.7%) and DENV-4 (92.3%). Seropositivity for DENV-1, DENV-2 and DENV-4 was maintained at 100% through months 3, 6 and 12. For DENV-3, 100% seropositivity was sustained at months 2, 3 and 12, with a slight decline to 91.7% at month 618. In parallel, our SRIP-based neutralization assay, performed using the same cohort, demonstrated 100% seropositivity for DENV-1, DENV-2 and DENV-4 from month 1 through months 2, 3, 6 and 12. For DENV-3, seropositivity remained at 100% from months 1 to 3, but declined to 93.3% at month 6 and further to 86.7% at month 12 (Table 1). We observed tetravalent responses in 100% of the participants from months 1 to 3, decreasing to 93.3% at month 6 and 86.7% at month 12 (Table 1). This decline may be attributed to lower nAb titers detected by the SRIP-based assay compared to those measured by the standard FRNT used in the KD-382 Phase I clinical trial18. While the SRIPs matched the genotypes of the KD-382 vaccine strains, they were derived from different viral isolates (GMTs at 12 months post-vaccination: DENV-1: 195, DENV-2: 121, DENV-3: 30, DENV-4: 146; Supplementary Table 1). In contrast, the FRNT used strain-matched wild-type viruses from the vaccine formulation (GMTs: DENV-1: 428, DENV-2: 710, DENV-3: 79, DENV-4: 152)18, which may explain the higher observed nAb titers due to strain-matched antigenic responses.

We evaluated the ability of KD-382 to induce cross-nAbs against 17 genetically diverse DENV genotypes. At 12 months post-vaccination, KD-382 induced broad and durable genotype-specific nAb responses. Notably, we observed strong nAb responses against DENV-1 Genotype II, DENV-2 Asian II and American genotypes, with GMTs of 232 (125–426), 532 (280–1009) and 232 (80–672), respectively (Fig. 2; Supplementary Table 1). Homologous genotypes included in the vaccine formulation – DENV-1 Genotype I, DENV-2 Asian I and DENV-4 Genotype I—elicited strong nAb responses, with GMTs of 192 (95–389), 121 (63–235), and 146 (60–356), respectively. In contrast, we observed lower nAb titers against all DENV-3 genotypes, with GMTs below 50. DENV-3 Genotype II, which is present in the vaccine formulation, demonstrated moderate neutralization (Supplementary Table 1). This is consistent with prior studies reporting comparatively weaker nAb responses to DENV-3 without necessarily compromising protection9,10. In NHPs, a single low dose of KD-382 (3 Log10 FFU per serotype) conferred protection against wild-type DENV-3 challenge, despite low vaccine-induced nAb titers and undetectable viremia following vaccination10. In the KD-382 Phase I clinical trial18, nAb titers for all four DENV serotypes showed no significant differences between single- and two-dose regimens, or between standard and low dose groups. This lack of a dose- or scheduled-dependent increase suggests that nAb responses may reach an early plateau, with even a single low dose providing sufficient immunological priming. These observations indicate that protection may not depend solely on the magnitude of vaccine-induced nAb responses or viral replication, but may instead involve broader immune mechanisms such as cell-mediated responses. Supporting this view, a human challenge study with other flaviviral vaccines—such as JEV and YFV—has demonstrated protection in the absence of detectable nAbs, attributed to robust and durable T cell responses19. Although the underlying mechanisms may differ, these findings reinforce the concept that effective vaccine-induced immunity is multifaceted and may not be exclusively reliant on high levels of circulating nAbs. Nevertheless, the reduced GMTs to DENV-3 should be interpreted with caution. As we were not able to perform antibody depletion experiments, we cannot definitively distinguish between homotypic and cross-reactive nAb responses. Further complexity is seen in the KD-382 Phase I clinical trial18, where vaccine-induced viremia was frequently detected for DENV-2 (100%) and DENV-4 (93.3%), aligning with strong nAb responses. DENV-1, despite limited viremia, elicited robust nab titers, while DENV-3 showed both low viremia and low nAb titers18. These findings suggest that vaccine-induced viral replication alone does not fully predict the magnitude or quality of the immune response. Rather, these findings underscore the possible importance of antigenic properties, host factors and the broader immune context in shaping protective immunity. Collectively, these findings indicate that a single low dose of KD-382 appears sufficient to prime protective immunity, with serotype-specific differences in response likely influenced by a combination of viral- and host-related factors.

Fig. 2: Genotype-specific longitudinal nAb titers elicited by a single low-dose administration of KD-382.
figure 2

For nAb titers (NT50: 50% neutralization) calculated to be <10, a value of 5 (assay’s limit of detection of 10 divided by 2) was assigned for analysis. Dashed lines indicate the assay’s limit of detection (<10). The nAb titer data are presented as geometric means with 95% CI.

Full size image

Given the risk of ADE in DENV infections, we evaluated whether KD-382-induced antibodies in flavivirus-naïve individuals could enhance viral infection in FcγR-expressing cells. Using BHK-21 cells stably expressing human FcγRIIA20, we assessed in vitro infection-enhancement across a range of serum dilutions. To characterize ADE dynamics, we calculated two key metrics: fold enhancement and peak enhancement titer (PET). Fold enhancement quantifies the relative increase in infection due to antibody-mediated uptake, expressed as the ratio of infection levels in the presence versus absence of serum. PET, on the other hand, refers to the serum dilution at which maximal enhancement is observed21, and may indicate the dilution range most likely to mediate ADE. While fold enhancement reflects the overall strength of enhancement in vitro, PET provides insight into the antibody concentrations at which enhancement is most physiologically relevant. Specifically, if peak enhancement occurs at high serum dilutions (i.e., low concentrations of enhancing antibodies), it may be less clinically relevant, as such low concentrations are rarely encountered in vivo. Conversely, PET values at low serum dilutions (i.e., high concentrations of enhancing antibodies) may signal a greater potential risk for ADE under physiological conditions (Supplementary Figure 1).

Our results showed that all 17 DENV genotypes tested exhibited infection-enhancement in vitro, with variations in the magnitude of enhancement across genotypes within each serotype. The highest mean peak fold enhancement was observed for DENV-2 Asian I (71.2 ± 35.9), Asian II (42.7 ± 20.5), and Cosmopolitan (45.3 ± 28.6) genotypes (Fig. 3). However, these increases were likely counterbalanced by robust nAb titers, as evidenced by high GMTs for these DENV-2 genotypes (Supplementary Table 1)—suggesting that neutralization predominated over enhancement in vaccinated individuals. In contrast, the lowest mean peak fold enhancement was observed in DENV-1 Genotype V (2.5 ± 1.1), DENV-2 American genotype (8.7 ± 5.0), and DENV-4 Genotype I (6.3 ± 4.0) (Fig. 3), consistent with high nAb titers against these strains (Supplementary Table 1). The observed variations in fold enhancement among the 17 DENV genotypes likely reflect the complex interplay between neutralization and ADE, as well as differences in the enhancement properties of the diverse DENV genotypes included in the SRIP panel.

Fig. 3: Overall genotype-specific peak fold enhancement levels of KD-382.
figure 3

Bars represent the mean ± SD of peak fold enhancement values. Genotypes included in the KD-382 vaccine formulation are shown in blue. Peak fold enhancement values are shown as means with SD.

Full size image

Regarding PET values, we observed a gradual decrease in PET starting over the 12-month study period (Fig. 4). However, the PET values remained outside the risk window for ADE22,23 (Fig. 4; Supplementary Table 2), indicating a low likelihood of ADE under the conditions tested. DENV-4 showed the most stability, with no significant fluctuation in PET values. In contrast, DENV-1 and DENV-2 exhibited significant declines in PET values at 6 and 3 months, respectively, but their PET values did not fall within the risk window for ADE. DENV-3 showed the lowest PET values among the other serotypes, and by the 12-month mark, DENV-3 had the lowest geometric mean PET value of 93 (76–113), though it still remained outside the risk window for ADE (Supplementary Table 2). Certain genotypes showed a gradual decline in PET values over time. Despite these declines, the stable peak fold enhancement levels observed between 2 and 12 months (Fig. 5; Supplementary Table 2) suggest that the nAb responses induced by a single low dose of KD-382 likely helps mitigate potential ADE. Additionally, serotype-specific PET values showed a moderately significant positive correlation with the corresponding GMTs (Supplementary Fig. 2), indicating that KD-382-induced neutralizing activity predominates over enhancing activity. This suggests a lower risk of ADE in vivo, as higher nAb titers are associated with reduced in vitro ADE, evidenced by high PET values (high serum dilution) that reflect lower concentration of enhancing antibodies. These findings highlight that a single low dose of KD-382 elicits a favorable immune profile skewed toward functional, protective immunity, with minimal enhancement-prone responses. Altogether, the data suggest a low theoretical risk of vaccine-induced ADE following administration of a single low dose of KD-382, sustained for at least 12 months.

Fig. 4: Enhancing antibody titers following a single low dose of KD-382.
figure 4

a Overall and serotype-specific longitudinal PET values. b Genotype-specific longitudinal PET values. PET values are presented as geometric means and 95% CI. PET values were calculated only for samples with peak fold enhancement levels ≥2.5, as determined using an ADE-negative control serum (from a DENV-naïve, DENV IgG-negative individual). Statistical significance for (a) was assessed using the non-parametric Kruskall–Wallis test, followed by pairwise comparisons of each time point from months 2 to 12 against month 1 as the reference group, using Dunn’s test. The ADE risk thresholds are depicted as shaded regions: threshold 1 (1:20–1:8022) and threshold 2 (≤1:4023), represented by dashed lines overlaid on a pink background. The fold enhancement cut-off at 2.5 was set based on the mean peak fold enhancement observed across all 17 DENV genotypes using the ADE-negative control serum.

Full size image
Fig. 5: Serotype-specific longitudinal peak fold enhancement levels of KD-382 over the 12-month observation period for each DENV serotype.
figure 5

Peak fold enhancement values are presented as means ± SD. Statistical significance was assessed using the non-parametric Kruskall–Wallis test, followed by pairwise comparisons of each time point from months 2 to 12 against month 1 as the reference group, using Dunn’s test.

Full size image

To our knowledge, this is the first report to apply the PET metric in evaluating a dengue vaccine candidate. While PET has previously been explored indirectly in natural infection settings22,23, we propose its potential value in assessing vaccine safety and waning immunity. Since enhancement appears to be an intrinsic property of DENV-specific antibodies—given that most human antibodies derived from primary and secondary DENV infections are capable of mediating ADE in vitro24—the critical determinant of ADE risk may lie not in enhancement per se, but in whether enhancement occurs at physiologically relevant antibody concentrations. This concept aligns with earlier findings by Kliks et al.21 who observed higher PET values in asymptomatic cases compared to hospitalized dengue patients, based on enhancement in primary human monocytes. In our study, PET values for all four DENV serotypes consistently exceeded the proposed ADE risk threshold (PET > 80)22, throughout the 12-month follow-up in flavivirus-naïve healthy adults. These findings suggest that a single low dose of KD-382 may confer protective benefit for at least 12 months in this population. However, a limitation is the absence of follow-up data beyond 12 months. Since severe dengue risk is thought to peak within two years of primary exposure, extended monitoring is essential to fully assess KD-382’s durability and safety with respect to ADE. In this context, a booster dose at 12 months may help sustain protective nAb levels and reduce the risk of breakthrough infections in dengue-naïve individuals. In dengue-immune individuals, the same regimen could enhance cross-nAb breadth and magnitude. Altogether, our findings support further evaluation of KD-382’s utility in both naïve and previously exposed populations with longer-term follow-up.

Another important consideration in interpreting PET values is the methodological difference between our study and previous work defining ADE risk threshold using inhibition enzyme-linked immunosorbent assay (ELISA)22. While inhibition ELISA correlates with hemagglutination inhibition25 and plaque reduction neutralization test results22, and measures antibodies targeting cross-reactive epitopes implicated in ADE in vitro26,27 and in vivo28, it does not assess functional outcomes of antibody-virus interactions. In contrast, our in vitro ADE assay evaluates antibody-mediated enhancement in FcγR-expressing cells, providing a more direct measure of functional activity. These methodological differences may explain inconsistencies across studies and may partly explain the challenges in linking ADE titers to severe dengue outcomes in cohort studies.

While ADE in vivo is primarily mediated by immune cells like monocytes and macrophages, we utilized BHK-21 cells stably expressing human FcγRIIA—an Fcγ receptor implicated in dengue pathogenesis. Although validation in immune cell models such as THP-1 is valuable, limited serum volume and the complexity of repeating the entire panel under multiple conditions currently preclude this. Our findings highlight the need for standardized, functionally based assays to more accurately assess ADE risk. In this context, measuring cross-reactive antibody function—without depleting type-specific antibodies – is particularly noteworthy as ADE is typically mediated by cross-reactive, non-neutralizing or sub-neutralizing antibodies. While we did not quantify neutralizing versus enhancing titers, observing in vitro enhancement at physiologically relevant concentrations underscores a critical safety consideration. Any dengue vaccine eliciting enhancement at such concentrations warrants close evaluation7. These findings highlight the importance of evaluating ADE potential in relation to cross-reactive antibody responses elicited by dengue vaccines.

Phase I clinical trials in flavivirus-naïve or dengue-naïve individuals for Dengvaxia (CYD-TDV), Qdenga (TAK003) and Butantan-DV (TV005) have demonstrated varying immunogenicity profiles. Qdenga, for instance, elicited high seropositivity rates 27 months after a single dose in dengue-naïve children and adolescents, with the strongest response against DENV-2 (100%; GMT: 356), followed by DENV-1 (90%; GMT: 73), DENV-4 (94%; GMT: 56), and DENV-3 (92%; GMT: 55)29. This bias toward DENV-2 is likely due to the DENV-2 backbone used for all four serotypes in Qdenga’s formulation. In contrast, a single-dose study of Dengvaxia showed lower immunogenicity in dengue-naïve children in the Philippines, with seropositivity rates of 28.7% (DENV-1), 29.4% (DENV-2), 21.3% (DENV-3), and 52.9% (DENV-4), and low GMTs (DENV-1 to DENV-4: 18, 20, 14 and 23, respectively) at 17 and 28 months post-vaccination30. One explanation for Dengvaxia’s suboptimal performance is the absence of DENV NS proteins in its chimeric YFV backbone, potentially compromising T cell priming and increasing breakthrough infection risk in vaccine recipients without prior DENV exposure. Butantan-DV demonstrated more promising results in flavivirus-naïve individuals, with single-dose TV005 administration yielding seropositivity rates of 92% (DENV-1), 97% (DENV-2), 97% (DENV-3), and 97% for (DENV-4) by day 90 post-vaccination and GMTs of 35 (DENV-1), 91 (DENV-2), 100 (DENV-3) and 205 (DENV-4)31. A controlled human infection study using TV005 also showed strong seropositivity rates of 75% (DENV-1), 100% (DENV-2), 95% (DENV-3), and 95% (DENV-4) by day 90 post-vaccination and GMTs of 42 (DENV-1), 159 (DENV-2), 71 (DENV-3) and 93 (DENV-4)32. By comparison, KD-382 demonstrated robust and durable immune responses in flavivirus-naïve individuals with a single-dose regimen. At 12 months post-vaccination, seropositivity remained high: 100% for DENV-1, DENV-2 and DENV-4, and 86.7% for DENV-3, with GMTs of 120, 128, 24 and 88, respectively. These findings underscore the durability and breadth of KD-382-induced antibodies relative to other major dengue vaccines.

A distinguishing feature of KD-382 is its formulation as a LATV derived from wild-type viruses for each serotype, unlike the chimeric constructs used in other dengue vaccines. Although T cell responses were not assessed in the Phase I clinical trial, the presence of serotype-specific NS proteins suggests potential for balanced cellular immunity—critical for protection in flavivirus-naïve individuals, who are at greater risk of ADE. All four dengue vaccines have utilized DENV strains mostly isolated in Southeast Asia between the 1960s and 1980s, raising concerns about mismatch with currently circulating genotypes. For instance, Dengvaxia’s reduced efficacy and ADE risk may partly stem from poor antibody match to these strains33, compounded by limited genotype-specific evaluation. In contrast, our study, evaluated KD-382 against a broad panel of DENV genotypes supporting its potential for both serotype and genotype cross-protection. Further investigation of T cell responses is essential to validate KD-382’s safety and efficacy as a single-dose dengue vaccine.

We acknowledge several limitations in this study. First, the small sample size limits the generalizability of the findings and warrants cautious interpretation of the results. Second, since the study focused on a low-immunity scenario following a single low-dose vaccination of KD-382, the conclusions may not directly apply to individuals receiving higher doses or multiple immunizations. Additionally, since our analysis was based on overall antibody responses without depletion of cross-reactive antibodies, it is difficult to precisely distinguish the contributions of type-specific versus cross-reactive antibodies to both neutralization and enhancement. Previous work by Aravinda de Silva and colleagues suggests that the presence of DENV type-specific nAbs—those directed specifically against a single serotype, independent of cross-reactive responses—may serve as a more reliable correlate of protection in dengue vaccine studies34,35,36. Type-specific antibody responses are thought to better reflect successful priming by each vaccine component and are less likely to mediate ADE compared to cross-reactive, lower-affinity antibodies. Despite these limitations, we believe that this focused functional analysis represents an important initial step toward understanding the potential ADE risk profile of KD-382 in dengue-naïve individuals and contributes valuable information for ongoing vaccine evaluation efforts.

In summary, this study shows that a single low dose of KD-382 can induce immune responses across multiple DENV serotypes and genotypes in flavivirus-naïve individuals. Although in vitro ADE was detected, it occurred primarily outside physiologically relevant concentrations. These results offer valuable preliminary insights into the immunogenicity and safety profile of KD-382 and highlight its potential as a tetravalent dengue vaccine candidate. Importantly, this study provides a balanced assessment of both neutralization and enhancement, contributing to ongoing efforts to develop dengue vaccines that address the complex interplay between immunity and ADE, especially in the context of DENV genetic diversity. Further studies, including clinical efficacy and T cell response evaluations, are needed to assess the suitability of KD-382 for broader use.

Methods

Ethics approval information

Ethical approval for this study was obtained from the Institutional Review Board of Nagasaki University (Approval number: 230615297). All serum samples were collected from participants who provided written informed consent as part of the Phase I clinical trial of KD-382 (ACTRN12618001027202).

Study design of the Phase I clinical trial

Serum samples analyzed in this study were obtained from healthy adult volunteers who participated in a Phase 1 clinical trial of the KD-382 dengue vaccine (ACTRN12618001027202), registered on June 19, 2018, and conducted by KM Biologics Co., Ltd. The Phase I clinical trial is publicly accessible at https://anzctr.org.au/Trial/Registration/TrialReview.aspx?id=375233&showOriginal=true&isReview=true. This randomized, placebo-controlled, double-blind, ascending-dose, single-center study was carried out in Australia, a non-dengue endemic country. To confirm dengue seronegativity, all participants were screened using FRNT against the four DENV serotypes. In addition, serostatus for other primary flaviviruses associated with human infection and disease—including Japanese encephalitis, Yellow Fever, West Nile, Zika and Kunjin—was assessed using ELISA. All participants were confirmed negative for these flaviviruses and were classified as flavivirus-naïve based on their negative results for dengue and other primary human-pathogenic flaviviruses. A total of 60 participants, aged 18–65 years, were enrolled between July 2018 and June 2020. The primary objective of the Phase I clinical trial was to evaluate the safety, tolerability, and immunogenicity of the KD-382 vaccine in flavivirus-naïve healthy adults. Immunogenicity was assessed by measuring seropositivity against wild-type parental viruses representing each DENV serotype. Seropositivity was defined as a post-vaccination FRNT50 titer ≥1:10 in participants whose baseline FRNT50 titer was <1:10. Participants were randomized to receive either the KD-382 vaccine—at a low dose (3 Log10 FFU per serotype) or standard dose (5 Log10 FFU per serotype)—or a placebo, administered subcutaneously as either a single dose or a two-dose regimen with a 4-week interval between doses. The Phase I clinical trial was conducted in two sequential parts. In Part 1, 30 participants were randomized in a 3:2:1 ratio to receive either a low dose of KD-382 as a single or two-dose regimen, or placebo. In Part 2, an additional 30 participants were randomized using the same ratio to receive either a standard dose or placebo.

Serum samples used in the current study

This exploratory study focused on a subset of participants from Part 1, specifically those in Sequence 1 (n = 15), who received a single low dose of KD-382. Participants in Sequence 2 (n = 10; two-dose recipients) and Sequence 3 (n = 5, placebo group) were excluded in this study. Serum samples from the 15 participants in Sequence 1 were collected at 1, 2, 3, 6, and 12 months post-vaccination. It is important to note that due to limited sample volume, particularly for the genotype-wide analysis, not all participants had serum samples available at every time point, with Month 1 (n = 12) and Month 2 (n = 7) being the most affected. The serum samples were used to assess the kinetics and durability of nAb responses, ADE activity, and DENV-specific IgM/IgG levels. Prior to analysis, all serum samples were heat-inactivated at 56 °C for 30 min to eliminate any residual virus and prevent potential interference in downstream assays.

KD-382 tetravalent vaccine description

KD-382 is a live attenuated tetravalent dengue vaccine composed of attenuated viruses from all four DENV serotypes: DENV-1, DENV-2, DENV-3 and DENV-4. The vaccine was formulated by combining monovalent virus solutions of each serotype. The parental strains for these viruses were sourced as follows: DENV1/03135 and DENV4/1036 were isolated from dengue fever patients, while DENV2/99345 and DENV3/16562 were isolated from dengue hemorrhagic fever patients9,10. To create the attenuated strains, DENV-1, DENV-2 and DENV-4 were serially passaged in certified primary dog kidney (PDK) cells, which were procured from the Netherlands. DENV-3, on the other hand, was attenuated through serial passage in PDK cells following an initial passage on primary green monkey kidney cells9,10.

Cell lines

Baby hamster kidney cells (BHK-21) were maintained in Eagle’s minimum essential medium (EMEM, Wako) supplemented with heat-inactivated 10% fetal bovine serum (FBS, BioWest) and 100 U/mL penicillin–streptomycin solution (Sigma-Aldrich). BHK-21 cells that stably express the human FcγRIIA20 were maintained in EMEM supplemented with heat-inactivated 10% FBS and 0.5 mg/mL neomycin (G418, PAA Laboratories). Human embryonic kidney 293T (HEK293T) cells were maintained in high glucose Dulbecco’s modified Eagle’s medium with l-glutamine (DMEM, Wako) supplemented with heat-inactivated 10% FBS, 1X MEM non-essential amino acids (Gibco) and 100 U/mL penicillin–streptomycin solution (Sigma-Aldrich). Cells were cultured at 37 °C in a 5% CO2 incubator.

Plasmid construction

The sub-genomic replicon plasmid from the YFV 17D vaccine strain (X03700) containing the nano-luciferase gene (pCMV-YFV-nluc-rep) and the expression plasmid for YFV 17D vaccine strain mature capsid consisting of 100 amino acids (pCAG-YF-C) were constructed as described previously16. To generate the expression plasmids for the signal, precursor membrane (prM) and envelope (E) of DENV-1 D1/Hu/Saitama/NIID100/2014 strain Genotype I (LC011945, Japan), DENV-1 99St-12A strain Genotype IV (MF094253, Philippines), DENV-2 00St22/2000 strain Asian II (AY786374, Philippines), DENV-2 36_DENV2 strain Cosmopolitan (MN083232, Sri Lanka) and DENV-3 Genotype I IDN2023 (PP587741, Indonesia), strains isolated from patients with dengue fever, were used. Viral RNAs were extracted from infected C6/36 cells, and reverse transcribed into cDNA. The cDNAs encoding the signal, prM and E were amplified by PCR, and the resulting fragments were cloned into the pCAGGS vector via in-fusion cloning technology (Takara Bio). The pCAGGS-based expression plasmid encoding the signal, prM and E of the other DENV genotypes and serotypes were constructed from synthetic DNA (GENEWIZ) based on the nucleotide sequence information available in GenBank (Supplementary Table 3). Plasmids were sequenced before use in the transfection experiments.

Production and titration of single-round infectious DENV particles

HEK293T cells grown in a 10-cm dish were co-transfected with three plasmids: 2.5 µg of replicon plasmid, 1.25 µg of capsid-expression plasmid, and 1.25 µg of prME-expression plasmid, using Polyethylenimine HCI Max, MW 40,000 (Polysciences) in Opti-MEM (Gibco) as described previously37. After 5–6 h, the culture medium was replaced with fresh medium. At 2 days post-transfection, the medium was replaced with complete medium supplemented with 10 mM HEPES buffer (Gibco). Standard SRIP preparations were harvested at 3 days post-transfection, were clarified through a 0.45 µM filter, and were frozen at −80 °C until further use. Titration of the harvested SRIPs was performed in 96-well plates using mock neutralization/ADE to determine the working dilution. SRIPs were serially diluted two-fold horizontally across the plate, combined with 1:1 (v/v) EMEM (Wako) supplemented with 2% FBS (Biowest), and incubated at 37 °C for one hour. Each plate included a cell control (no SRIP), and each SRIP sample was assayed in four replicates. The SRIP dilutions were inoculated onto plates containing cell monolayers and were incubated at 37 °C in 5% CO2 for 5 h. Fresh medium was then added, and the plates were further incubated at 37 °C in 5% CO2 for 2 days. The luciferase activities of SRIP-infected cells were detected at 2 days post-inoculation using the Nano-Glo® Luciferase Assay System (Promega) and were measured with an integration time of 1 s in a BioTek Synergy H1 Multimode Reader (Agilent Technologies). An acceptance criterion of ≥100-fold difference between the SRIP control and the cell control in all four replicates of the SRIP tested (e.g. 105 relative luminescence unit (RLU) for the SRIP control and 103 RLU for the cell control) was set; therefore, the maximum dilution of the SRIP that met this criterion was selected as the working dilution for the neutralization and ADE assays.

Neutralization and ADE assays using single-round infectious DENV particles

The neutralization experiments were conducted using BHK-21 cells, whereas ADE experiments were conducted using FcγRIIA-expressing BHK-21 cells. Serum samples were heat-inactivated at 56 °C for 30 min. Cells were seeded in 96-well plates at a density of 7.0–8.0 × 104 cells per well overnight. Serum samples were serially diluted four-fold with EMEM (Wako) supplemented with 2% FBS (Biowest), mixed with SRIPs at a 1:1 ratio, then were incubated at 37 °C for one hour. The SRIP-serum mixtures were inoculated onto plates containing cell monolayers and were incubated at 37 °C in 5% CO2 for five hours. Fresh medium was added, and plates were further incubated at 37 °C in 5% CO2 for two days. Each plate included a SRIP control (no serum) and cell control (no SRIP, no serum), with each serum sample assayed in duplicate technical replicates. The luciferase activities of SRIP-infected cells were detected at 2 days post-inoculation using the Nano-Glo® Luciferase Assay System (Promega) and were measured with an integration time of 1 s in a BioTek Synergy H1 Multimode Reader (Agilent Technologies). The nAb titer was expressed as the serum dilution that reduced ≥50% luciferase signal. The reduction of luciferase signal was calculated by the following formula:

$$\% \,{\text{luciferase}}\,{\text{reduction}}=\left(1-\frac{{{\text{mean}}\,{\text{RLU}}}_{\text{t}}-{\text{mean}}\,{\text{RLU}}_{\text{c}}}{{\text{mean}}\,{\text{RLU}}_{\text{s}}-{\text{mean}}\,{\text{RLU}}_{\text{c}}}\right)\times 100 \%$$

The fold enhancement ratio was calculated by the following formula:

$${\text{fold}}\,{\text{enhancement}}=\,\frac{{\text{mean}}\,{\text{RLU}}_{\text{t}}-{{\text{mean}}\,{\text{RLU}}}_{\text{c}}}{{\text{mean}}\,{\text{RLU}}_{\text{s}}-{\text{mean}}\,{\text{RLU}}_{\text{c}}}$$

RLU is the relative luminescence unit measured corresponding to the luciferase activity in SRIP-infected cells. RLUt is the RLU of the test serum sample, RLUs is the RLU of the SRIP control, while RLUc is the RLU of the cell control. The peak enhancement titer (PET) was determined by plotting the fold enhancement value on the y-axis and log-reciprocal serum dilution on the x-axis. A Gaussian distribution curve was used to fit the ADE curve using GraphPad Prism 10 (GraphPad Software). The data point corresponding to the amplitude was used to derive the log-reciprocal serum dilution and was assigned as the PET.

Detection of DENV-specific IgM using IgM-capture ELISA

The in-house DENV IgM-capture ELISA was performed according to the protocol described by Bundo and Igarashi13. 96-well microplates were coated with 100 μL of anti-human (μ-chain specific) goat IgG (5.5 μg/100 μL) (Cappel ICN Pharmaceuticals) diluted in ELISA coating buffer (0.05 M carbonate-bicarbonate buffer, pH 9.6, containing 0.02% sodium azide), with all wells except the blank wells receiving the coating solution. Plates were incubated at 4 °C overnight to allow for antigen adherence. After incubation, the wells were blocked with 100 μL of BlockAce (Yukijirushi) at room temperature for one hour, except for the blank wells. Following blocking, the plates were washed three times with phosphate buffered saline containing 0.05% Tween-20 (PBS-T). Serum samples, including positive and negative controls, were diluted 1:100 in PBS-T and 100 μL aliquots were added to duplicate wells. The plates were incubated at 37 °C for one hour, after which the serum was removed and the plates were washed as previously described. To detect bound IgM, 100 μL of a tetravalent DENV antigen (25 ELISA units per serotype) was added to each well and incubated at 37 °C for one hour. After washing, the plates were allowed to react with 100 μL of horseradish peroxidase (HRP)-conjugated anti-DENV, mouse monoclonal antibody (12D11/7E8)38, diluted 1:3000, for one hour at 37 °C. Following another wash, the color was developed by adding 100 μL of substrate solution containing 5 mg o-phenylenediamine dihydrochloride (OPD) (Sigma Chemical) and 0.03% hydrogen peroxide in 10 mL of 0.05 M citrate–phosphate buffer, pH 5.0. The reaction was allowed to proceed for one hour in the dark at room temperature, then terminated with 100 µL of 1 N sulfuric acid. The optical density (OD) was measured at 492 nm using the BioTek Synergy H1 Multimode Reader (Agilent Technologies). A positive result was determined if the P/N ratio (positive control OD492/negative control OD492) was ≥2.0.

Detection of DENV-specific IgG using indirect ELISA

In this study, an in-house DENV IgG indirect ELISA was employed, based on the method described by Inoue et al.14, to assess DENV-specific IgG levels. The ELISA was performed using 96-well microplates coated with 100 µL of antigen (250 ng of each serotype/100 µL/well), which was diluted in ELISA coating buffer, except for the blank wells. The plates were incubated at 4 °C overnight to ensure proper antigen adsorption. Following coating, the wells were blocked with 100 µL of BlockAce (Yukijirushi) at room temperature for 1 h to reduce non-specific binding. After blocking, the plates were washed three times with PBS-T to remove unbound proteins and blocking agents. Serum samples were diluted at 1:100 in PBS-T with 10% BlockAce (Yukijirushi) and added in duplicate to the antigen-coated wells. Dengue-positive control serum, known to contain antibodies specific to DENV, was included on each plate for comparison. The plates were incubated at 37 °C for one hour to allow specific binding of antibodies. After this incubation, the wells were washed again, and 100 µL of 1:30,000 diluted HRP-conjugated goat anti-human IgG (American Qualex) in PBS-T with 10% BlockAce (Yukijirushi) was added to each well. This secondary antibody binding was allowed to occur at 37 °C for 1 h. The peroxidase reaction was initiated by adding 100 µL of a substrate solution containing 5 mg of OPD (Sigma Chemical) and 0.03% hydrogen peroxide in 10 mL of 0.05 M citrate phosphate buffer (pH 5.0). Plates were incubated at room temperature for 15–30 min in the dark to develop the color. The reaction was then stopped by adding 100 µL of 1 N sulfuric acid to each well. The OD was measured at 492 nm using the BioTek Synergy H1 Multimode Reader (Agilent Technologies). A standard curve was constructed using OD492 values from dengue-positive control serum, starting with a 1:1000 dilution and performing serial two-fold dilutions down to 1:212. The IgG titers of the serum samples were determined based on this standard curve. A sample IgG titer of ≥1:3000 was considered DENV IgG-positive, with the cut-off value for positive IgG defined as 1:1000 ± 3 standard deviations.

Statistical analysis and graphical representations

Statistical analyses were performed using GraphPad Prism 10 (GraphPad Software), with a 5% level of significance and two-tailed p-values. Missing data were handled by listwise deletion, and any observations with missing values were excluded from the analysis. The amount of missing data was minimal and assumed to be missing completely at random. To assess monthly differences in nAb titers and PET values, a Kruskal–Wallis test was conducted, followed by post hoc pairwise comparisons using Dunn’s test with Bonferroni correction for multiple comparisons. To assess overall genotype-specific peak fold enhancement throughout the 12-month observation period, an ordinary one-way ANOVA was performed, with multiple comparisons among the 17 genotypes corrected using the Tukey test. The detailed specification of the statistical analyses used for each figure is provided in the figure legends of the corresponding figures. Figures 1–3 and 5 were prepared using Graphpad Prism 10, and Fig. 4 was prepared using BioRender (https://www.biorender.com) for publication.